Hello, Im running a did model with individual fixed effects on two waves of a longitudinal survey. My treatment is the death of a HH member. I have a 7-year gap between wave 1 and wave 2. So the death of the individual can happen at any point in time in the gap. I also have data on how many years it has been since the individual passed away(point of reference wave 2). But the outcome variable is only available at two time periods ie at the year when wave 1 took place and when wave 2 took place.
1)So here do I have to be wary of treatment heterogeneity?
2) I have individual fixed effects so, for the group variable do I just hadd the individual ID or should I take it a level higher say HH ID.
3)Or should I use the dynamic verison
1)So here do I have to be wary of treatment heterogeneity?
2) I have individual fixed effects so, for the group variable do I just hadd the individual ID or should I take it a level higher say HH ID.
Code:
did_multiplegt outcome IndividualID WAVE treated, cluster(PSUID) controls (X)
Code:
did_multiplegt outcome IndividualID WAVE treated, robust_dynamic dynamic(1) breps(100) cluster(PSUID) controls (X)
Comment